多源干扰下侧扫声呐图像复原的综合校正研究

Research on comprehensive correction of side scan sonar image multiplication under multi-source interference

为了减少因海洋环境以及侧扫声呐工作机理等干扰所造成的声呐图像畸变,提出多源干扰下侧扫声呐图像复原的综合校正方法。针对辐射图像明暗区域差异明显、轮廓特征模糊等问题,基于双伽马函数和Ret-inex理论设计辐射均衡化和轮廓清晰化算子,实现了图像辐射校正;针对环境干扰及工作机理所造成的图像水柱区及图像压缩等问题,设计了海底点检测因子并通过获取实际平距实现斜距校正;针对AUV速度波动造成的图像异常拖拽和压缩等问题,基于经纬度设计航行速度,并通过校正比例系数复原图像块来实现速度校正。实验测试结果表明,相较于其它方法,文中综合校正方法不仅使校正图像具有均衡性好、特征轮廓清晰、水柱区基本消除和两舫图像衔接平滑的优点,而且SNR指标平均提升了6.060%、ENT指标平均降低了16.583%、SD指标平均降低了49.904%、ENL指标平均增加了78.347%,有效实现了侧扫声呐图像的复原。

In order to reduce sonar image distortion caused by interference from the marine environment and the work-ing mechanism of side scan sonar,a comprehensive correction method for side-scan sonar image restoration under multi-source interference is proposed.In order to solve the problems of obvious differences between light and dark areas of radi-ation images and blurred contour features,radiation equalization and contour clearing operators were designed based on the double gamma function and Retinex theory to achieve image radiation correction;In order to solve the problems of image water column area and image compression caused by environmental interference and working mechanism,the seabed point detection factor was designed and the slant distance correction was realized by obtaining the actual horizontal distance;In or-der to solve the problems of abnormal image dragging and compression caused by AUV speed fluctuation,the navigation speed was designed based on longitude and latitude,and the speed correction was achieved by restoring the image blocks by correcting the scale coefficient.Experimental test results show that compared with other methods,the comprehensive correc-tion method in this article not only makes the corrected image have the advantages of good balance,clear feature contours,basic elimination of water column areas and smooth connection of images on both sides,but also improves the SNR index by an average of 6.060%.The ENT index decreased by 16.583%on average,the SD index decreased by 49.904%on average,and the ENL index increased by 78.347%on average,effectively achieving the restoration of side scan sonar images.